1) Unless you're getting pay-site data, the EC only comes in time steps of 24hr. Since it releases runs every 12 hours, you're not getting to view the same time stamp from run to run. Thus, in this case, it very well could drop to the previous run's value in between tau times.
Right, that makes sense. I remember looking at GFS temperature maps and not understanding why there were these huge fluctuations every 12 hours. It took me a while to realize that this had to do with the difference between night and day!
I don't mind feeling stupid, as that means I'm catching myself. It's worse being stupid and not noticing it. On the other hand, ignorance is bliss.
But does the same apply to SLP? Ie, the fluctuations between 12-hour runs. I thought that maybe it had to do with models struggling to forecast an intense cyclone in June.
2) The point you make about data reliability at D9/D10 can't be stressed enough. Deterministic runs lose skill rapidly after D5. It's imperative to all who post here that the ensemble runs are posted instead. The only exception is when the deterministic run is supported by the ensemble, in which case it probably has some value. Ensemble means are specifically designed to reduce error in the medium-long range by means of averaging and dampening outlier solutions. In either case, if a deterministic run is posted past D5 (120hr) -- it should have it's accompanying ensemble posted as well OR at the very least have an agreeing deterministic run posted with it (e.g. those cases where the GFS and EC agree at those time ranges). That is -- if we're to take it seriously.
I wanted to ask a couple of stupid questions about ensembles, but then decided to look it up myself and found this
informative PDF from ECMWF. Thanks for putting me on this track, Csnavywx.
I will try and embed the EPS forecast (based on ECMWF ensemble) in my daily routine, but here I'll keep reporting on the D1-D6 forecast from the deterministic run, especially when it concerns cyclones, as I'm mostly interested in how low the SLP could go (and an ensemble will probably conservatively muffle the number).
But as a test and educative experience, I will make a side-by-side comparison of the D1-D6 forecasts of the ECMWF deterministic run, the EPS ensemble run, and the D0 images for the coming 6 days.
Here are both the deterministic and ensemble run for the coming 6 days. The ECMWF forecast (00Z run) now has 967 hPa for D4 (ensemble has 976 hPa):